Background β-Antagonists have recently been proposed for the treatment of chronic asthma; however, concerns regarding risk of acute bronchoconstriction in clinical trials remain.
Objective To determine the frequency of oral β-blocker prescribing in patients with asthma and associations with severe asthma exacerbations requiring oral steroids in patients with active asthma defined by prior asthma-related medication use.
Methods Patients with asthma registered on 31 March 2007 and all asthma-related medications from the preceding 2 years were identified from anonymised clinical data from one-third of Scottish general practices. The main outcome measures were the relative incidence of patients with active asthma receiving oral steroids following a new oral β-antagonist prescription.
Results Of the 53 994 adult patients identified with asthma, 1171 (2.2%, 95% CI 2.10% to 2.30%) patients were prescribed an oral β-antagonist. 971 patients (82.9%, 95% CI 80.7% to 85.0%) had active asthma, with 52.7% on preventer medication. Of the 869 active asthma patients with sufficient follow-up, the average number of patients prescribed rescue oral steroids at baseline was 6.6 (0.8%) patients every 2 weeks. Rescue steroids were prescribed to six patients (0.7%) in the first 2 weeks and to eight patients (0.9%) in the second 2 weeks following the new oral β-antagonist (incidence rate ratio (IRR) 0.92, 95% CI 0.40 to 2.12; and IRR 1.24, 95% CI 0.60 to 2.58, respectively). No significant difference was found following stratification for β-antagonist selectivity.
Conclusion These results suggest that prescribing new oral β-blockers for the purpose of investigating potentially beneficial effects of chronic treatment would not lead to large increases in patients treated with oral steroids acutely in general practice.
Statistics from Altmetric.com
Asthma is a common chronic disease with a high prevalence rate reported in Scotland (18.4%), compared with England (15.3%), New Zealand (15.1%), Australia (14.7%) and Canada (14.1%).1 Short-acting inhaled β2-agonists (SABAs) have a longstanding central role as the primary treatment for acute relief of asthma symptoms. More recently, long-acting β2-agonists (LABAs) have also been recommended for use as add-on controller therapy to inhaled corticosteroids (ICS) in patients requiring regular reliever therapy at step 3 of the guidelines.2 3 Recent guidance from the Food and Drug Administration (FDA) highlighted concerns regarding the long-term safety of LABAs and this has led to a re-appraisal of their role as add-on controller therapy to inhaled steroids.4 In contrast, β-adrenergic antagonists or β-blockers are typically contraindicated in asthma due to the potential to induce opposing effects, namely acute bronchoconstriction.5 6
The effects of β2-adrenoceptor antagonists have recently received close attention as studies have suggested that chronic treatment with the non-selective β2-adrenoceptor antagonist nadolol may lead to reductions in airway hyper-responsiveness, inflammation and upregulation of β2-adrenoceptors in the asthmatic mouse model.7–9 In a subsequent open label study involving steroid-naïve patients with mild asthma, chronic dosing with the non-selective β-blocker nadolol over 12 weeks caused a significant reduction in airway hyper-responsiveness to methacholine challenge along with a small residual fall in forced expiratory volume in 1 s (FEV1) after an initial greater fall on first dose exposure.10 So-called cardioselective β-blockers, such as atenolol and bisoprolol, demonstrate greater affinities for β1- than β2-adrenoceptors, although the ratios of β1:β2 antagonism show marked variation between different agents.11 The effects of cardioselective β-blocker exposure in those with mild to moderate asthma have been investigated using a meta-analysis of randomised, blinded placebo-controlled trials.12 Single dosing with cardioselective β-blockers produced a significant reduction in FEV1 of 7.46% without affecting symptoms, while FEV1 was not significantly reduced by chronic dosing. In addition, a significant increase in subsequent β-agonist response was seen upon chronic dosing, indicating that β2-receptor upregulation may occur. This could explain how the beneficial effect of β-blockers in asthma may only be seen upon chronic dosing, as is the case with heart failure.
In a commentary on the therapeutic potential of β-blockers for asthma, the issue was raised that it would be important to evaluate the potential adverse effects of β-blockers on exacerbations of asthma from real-life prescribing data in patients who were already co-prescribed β-agonists and antagonists.13 Such prescribing is not uncommon, with a previous analysis of prescriptions issued in primary care in the Tayside region of Scotland demonstrating that ∼1.7% of patients receiving a β-blocker prescription had received a β2-agonist in the preceding 30 days.14 We wished to know how patients with asthma tolerated oral cardioselective and non-selective β-blockers in real life, especially in terms of precipitating severe exacerbations in order to inform the likely risk of performing further prospective randomised controlled trials to explore the therapeutic potential of β-blockers in asthma.
The aim of the present study was to examine the prevalence of oral β-blocker prescribing in a large cohort of people with asthma, and to determine whether the initial oral β-blocker prescription was associated with large increases in the number of patients receiving treatment with oral steroids for severe exacerbations of asthma.
Anonymised clinical data were obtained from the Primary Care Clinical Informatics Unit at the University of Aberdeen. The data extract was from 310 of the 1031 Scottish general practices with 1.76 million registered patients who are representative of the wider population in terms of age, sex and deprivation.15 Analysis of anonymised clinical data conformed to the Primary Care Clinical Informatics Unit research governance process.
Data were extracted for all patients with an asthma Read Code permanently registered on 31 March 2007 and included patient age, sex, smoking status, additional diagnosis of chronic obstructive pulmonary disease (COPD), and all asthma-related medications prescribed in the previous 2 years including SABAs, LABAs, ICS, leukotriene antagonists, theophyllines, oral steroids and β-blockers.
For patients aged 18–40, cases of asthma were defined as adult patients with a Read Code for asthma and no Read Code for COPD, defined by the Quality and Outcomes Framework (QOF) business rules as part of the new General Medical Services (GMS) contract for general practice in the UK. To ensure exclusion of patients with undiagnosed COPD, patients aged >40 were included only if they were additionally recorded as ‘never smoked’ according to their QOF Read Code for smoking status.
The main outcome measure was the incidence of patients receiving treatment with oral steroids in consecutive 2-weekly intervals. Due to the variation in oral steroid use and potential for confounding, analysis was performed using two separate oral steroid outcome measures. The main outcome measure was the incidence of patients receiving rescue oral steroid, defined as a total prednisolone prescription dose between 35 and 350 mg and a prescription duration of <2 weeks calculated using quantity, dose and frequency information on the prescription. The secondary outcome measure was the incidence of patients prescribed any oral steroid.
New β-blocker prescriptions in patients with active asthma
A new oral β-blocker prescription was defined as a prescription for an oral β-blocker with no previous prescription in the previous 140 days or longer. Active asthma was defined using prescriptions of asthma-related medication occurring between 1 April 2005 and the date of the new β-blocker prescription, including prescriptions of SABAs, LABAs, ICS, combination inhalers (LABA/ICS), theophylline and leukotriene antagonists. To determine the effects on oral steroid prescribing, patients were required to have a follow-up period of at least 112 days after a new oral β-blocker prescription.
British Thoracic Society/Scottish Intercollegiate Guideline Network (BTS) treatment step
The BTS guidelines for the management of asthma are UK national-level guidelines aimed to help both primary and secondary care clinicians in the management of asthma.2 They advocate a stepwise approach to the treatment of asthma in the UK. The BTS treatment step was calculated using prescriptions issued in the 84 days prior to the new β-blocker. These were defined as step 1 (SABA use only), step 2 (ICS ± SABA use only), step 3 (step 2 including add-on therapy with a LABA, leukotriene or theophylline) and step 4 (step 3 with high-dose ICS or with the addition of a fourth drug). High-dose ICS was defined as an inhaled ICS dose of 1000 μg beclomethasone equivalent or greater per day.
Binary logistic regression was performed using simple contrasting in order to determine the effects of β-antagonist prescribing between gender and age. Poisson regression with a generalised estimating equation (GEE) approach was used to determine incidence rate ratios (IRRs) of patients prescribed oral steroids between time periods around the first β-blocker prescription only. A model-based estimator using an autoregressive (AR1) correlation matrix was applied. The significance threshold was 5% and analysis was performed in SPSS version 17.0. The model was inspected for overdispersion by assessing the ratio of model χ2 to degrees of freedom and adjusted for BTS treatment step severity and season at the time of β-blocker initiation. The immediate 2 week pre-exposure period (days −14 to −1) was excluded from the analysis to account for event-dependent exposures. In this regard, patients receiving treatment for an exacerbation of asthma are less likely to be prescribed a β-blocker at the same time and would result in an exaggerated relative incidence.
A total of 53 994 adult patients with asthma were identified using our definition, of which 1171 patients (2.2%, 95% CI 2.10% to 2.30%) were issued a total of 6263 oral β-blocker prescriptions. Of the 1171 patients prescribed an oral β-blocker, 392 patients (33.5%, 95% CI 30.8% to 36.2%) received an average of 3.8 prescriptions (95% CI 3.4 to 4.1) for non-selective β-blockers compared with 779 patients (66.5%, 95% CI 63.8% to 69.2%) who received an average of 6.1 prescriptions (95% CI 5.8 to 6.3) for cardioselective β-blockers. Cardioselective β-blockers accounted for 4739 (75.7%) of all prescriptions issued.
Both age and gender had significant impacts on oral β-blocker prescribing in patients defined with asthma. The majority of patients prescribed a non-selective β-blocker (53.2%) were under the age of 40 while the majority of patients prescribed cardioselective β-blockers (91.5%) were ≥40 years of age. The overall odds of being prescribed any class of β-blocker increased incrementally with age (table 1).
Of the 779 patients prescribed a cardioselective β-blocker, the most commonly prescribed drugs were atenolol (62.9%) and bisoprolol (29.8%), with a mean daily starting dose of 51.5 mg for atenolol (95% CI 49.3 to 53.7 mg) and 5.3 mg for bisoprolol (95% CI 4.8 to 5.8 mg) (table 2). Of the 392 patients prescribed non-selective β-blockers, the most commonly prescribed drug was propranolol (83.2%), with a mean daily starting dose of 74.0 mg (95% CI 69.2 to 78.8 mg). Of the 1171 patients prescribed an oral β-blocker, 909 patients (77.6%) received more than one β-blocker prescription (figure 1).
Oral steroid prescribing in patients with active asthma and β-blockers
Of the 1171 patients prescribed an oral β-blocker, 847 patients (72.3%) received prescriptions for SABAs and 614 patients (63.2%) prescriptions for ICS before the new β-blocker (table 3). In total 971 patients were identified as having active asthma prior to the new β-blocker. The severity of asthma as defined by BTS treatment step for all asthma patients issued a new β-blocker and specifically for active asthma patients issued a new β-blocker is shown (table 4). Of all asthma patients receiving a β-blocker, 43.7% were at BTS treatment step 2 or greater compared with 52.7% of patients with active asthma.
Of the 971 patients with active asthma, 869 (89.5%) had follow-up of at least 112 days following the new oral β-blocker. The 2-weekly incidence of patients with active asthma treated with rescue oral steroids was 6.6 patients (0.8%) for baseline, 6 patients (0.7%, IRR 0.92, 95% CI 0.40 to 2.12) for the first 2 weeks, 8 patients (0.9%, IRR 1.24, 95% CI 0.60 to 2.58) for weeks 2–4 and 3.8 patients (0.4%, IRR 0.84, 95% CI 0.55 to 1.27) for weeks 4–8 following the new oral β-blocker prescription table 5. The 2-weekly incidence of patients with active asthma treated with any oral steroid for the equivalent time periods was 11.7 patients (1.3%), 19 patients (2.2%, IRR 1.63, 95% CI 1.00 to 2.65), 13 patients (1.5%, IRR 1.12, 95% CI 0.63 to 1.99) and 11.3 patients (1.3%, IRR 0.97, 95% CI 0.72 to 1.32), respectively. Figure 2 shows the proportion of patients with active asthma requiring oral steroid treatment before and after the new oral β-blocker prescription, and no major increase in oral steroid prescribing was seen immediately after the new β-blocker. No statistical difference was found in the proportion of patients prescribed oral steroids in the weeks following a new β-blocker prescription (table 5).
The main concern when designing studies to examine the potential therapeutic use of β-blockers in asthma remains the risk of inducing acute bronchoconstriction. It is likely that any benefit of β-blockers in asthma will only be seen upon chronic dosing, perhaps after an initial period of dose titration. Similar concerns existed for the use of β-blocker therapy in heart failure patients which can safely be initiated and is highly effective.16 The result of this population-based study showed that 1171 patients (2.2%) of 53 994 patients identified with asthma in the Scottish population were prescribed oral β-blocker therapy within a 2 year period. Selective β-blockers were prescribed more often in older patients who are more likely to have compelling indications for β-blockade such as ischaemic heart disease and heart failure. Female patients were more likely to be prescribed a non-selective β-blocker than males. This may represent the increased prevalence of conditions associated with non-selective β-blocker treatment in females such as anxiety, migraine, thyrotoxicosis and pre-eclampsia. Although reasons for prescribing at the level of primary care remain heterogeneous and complex, it is possible that this observation is partly due to the exclusion of patients from our population over the age of 40 with an exposure to smoking. Misclassification of patients with COPD would increase the observed incidence of β-blocker prescribing as current recommendations state that cardioselective β-blockers can be safely used in patients with COPD.17 Smoking status was recorded within 3 years of the date of extraction in >98% of patients over the age of 40. As such, we used patients over the age of 40 who had never smoked in order to reduce misclassification of COPD in our asthma population. Despite this, we were unable to exclude patients with non-smoking-related fixed airflow obstruction
Oral β-blocker therapy is currently stated to be contraindicated in asthma due to the potential to induce acute bronchoconstriction, and the Committee for Safety on Medicines in the UK recommends that cardioselective β-blockers should only be initiated with extreme caution under specialist supervision if no alternative is available.18 However, significant numbers of patients with asthma were prescribed oral β-blocker therapy in the community, with a third of our patients receiving a non-selective β-blocker. Irrespective of β-blocker selectivity, many patients received more than one prescription. Despite these safety concerns, the number of patients with active asthma prescribed oral steroids was remarkably low following the initial oral β-blocker prescription, ranging from 0.7% to 2.3% of patients. Two indicators of risk were chosen for comparison, namely patients with active asthma receiving rescue oral steroid prescriptions and any oral steroid prescriptions. This was done due to the wide variation and indication for oral steroid prescribing and to address the confounding effect this would have on masking any observed risk following initiation of β-blocker therapy. No statistically significant change in the number of patients requiring oral steroids related to new oral β-blocker prescribing was identified despite the fact that it takes up to 4 weeks for β2-receptor function to be restored back to normal following a non-selective β-blocker, possibly due to upregulation and associated enhanced coupling of G protein adenylyl cyclase.19 20 However the CIs are relatively wide and a small increase cannot be excluded.
Given UK patterns of oral β-blocker use, it is unsurprising that the most commonly prescribed cardioselective and non-selective β-blockers were atenolol and propranolol, respectively. However, the variation in daily starting prescription dose of new oral β-blockers demonstrates that patients with asthma prescribed β-blockers in this study did not always start with a period of dose titration. We have demonstrated that oral β-blocker therapy was prescribed to patients with active asthma of varying severity, with over half of new β-blocker prescriptions being issued to patients at BTS step 2 or greater. This observation means that the results are unlikely to be due to confounding by indication, in which only patients with stable well-controlled asthma requiring little medication are exposed to β-blocker therapy and hence have less need for rescue oral steroids. All patients in the main analysis received a new prescription for an oral β-blocker, and all had evidence of active asthma before commencement of the β-blocker in terms of a recent prior asthma-related prescription. With half of patients on BTS step 2 or higher, one would expect oral β-blocker prescribing in this population to be potentially even more hazardous.
The fact that patients prescribed a new oral β-blocker in this observational study potentially did not receive a period of dose titration means it may be possible to address any small risk by appropriate dose titration upon commencement of β-blocker therapy. In contrast, it is likely that any clinical trial looking at the potential benefits of β-blockers in asthma would start with this period of dose titration in order to identify patients whose symptoms are exacerbated at an earlier stage.
This study has several limitations. First, routine health data can be subject to missing data. However, the Primary Care Clinical Informatics Unit at the University of Aberdeen collects these data as part of the Scottish Programme for Improving Clinical Effectiveness (SPICE). The data that are provided are the complete historical patient record which includes all prescriptions issued by computer from general practice. As such, we have no knowledge regarding dispensed medication or patient compliance.
Secondly, the database contains anonymised data and therefore it was not possible to link to hospitalisation or mortality registers in order to identify whether any asthma patients prescribed oral β-blocker therapy received treatment in secondary care. This number, however, is likely to be small compared with the number of rescue steroids prescribed in general practice. The National Health Service (NHS) provides free access to primary care for all citizens in Scotland and patients register with a general practice in order to access this care. All prescriptions, including those recommended by secondary care services, are issued via primary care, with the exception of short courses of medication provided at the point of hospital discharge. The Scottish Public Health Observatory provides annual statistics on asthma for Scotland including asthma prevalence and incidence of hospital admissions.21 Based on this data, the incidence of hospital admission for asthma was 11.8 per 1000 asthma patients per year, compared with an incidence of 180.2 rescue steroids per 1000 asthma patients per year from our data. Therefore, prescribing during hospital admissions for asthma is estimated to account for ∼6% of all prescribing.
Given the current interest in the use of β-blockers as potential therapeutic agents for patients with asthma, these initial findings appear to provide some reassurance that co-prescribing of β2-agonists in patients with active asthma does not appear to result in a significant deterioration in asthma control requiring oral steroids in the vast majority of patients, although a small effect cannot be excluded. Further studies are now underway to evaluate the degree of airway hyper-responsiveness in such patients who have been taking β-blockers as well as their response to rescue therapy with β2-agonists.
A total of 2.2% of adult patients defined with asthma were prescribed oral β-blocker therapy which included prescriptions of both cardioselective and non-selective β-blockers in a 2 year observation period. Despite variation in starting dose and severity of asthma prior to the initial β-blocker, no large increase in the number of patients with active asthma treated in general practice for severe asthma exacerbations requiring oral steroids was observed in what would be considered to be the most high risk period immediately following the new oral β-blocker prescription, although a small increase in risk cannot be excluded.
Funding The study was funded by NHS Quality Improvement Scotland. NHS Education for Scotland provided funding for the role of DRM. Neither organisation had any influence on the analysis or publication of the work.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.